Many octave instances

[Bart Verleye]

Dear Octave developers,

We try to run about 8000 octave instances at the same time on our 
cluster. Running a single job goes perfectly, and takes about 15min. 
Running up to 300 jobs goes well too. However, more than that, and the 
program hangs.

From a little debugging, we think the cause is that every octave 
instance tries to read the directory it was called from, looking for .m 
files and functions. The directory contains quite a lot of files, and if 
8k instances try to read the same file(s), concurrency issues come up 
(we use gpfs).

Do you think this explanation makes sense? Is there any way I can avoid 
octave from reading the files in that directory?

Thanks and kind regards,
Bart

--
Dr. Bart Verleye
Centre for e-Research
Level G, Room 409-G21
24 SYMONDS ST
Auckland 1010
New Zealand
+64 (0) 9 923 9740 ext 89740